Agentic AI Chatbot Integration Support
Budget / SalaryHourly project
TypeFreelance project
LocationRemote
Posted1 hour ago
I have an in-flight chatbot project and I now need an agentic AI engineer who can take ownership of the integration layer. The heaviest lift is the AI side: wiring Langchain and LangGraph together, then connecting them to Azure AI Search so that embeddings, vector search and full RAG workflows run smoothly.
Every call the bot makes—GET, POST and PUT—must pass through a clean Python requests wrapper that you’ll write, document and unit-test. All code will live in a GitHub repo; I expect you to create concise branches, open pull-requests and keep the CI pipeline green.
What I’ll hand you:
• Current codebase with a minimal working chatbot
• Azure resources and keys
• Access to the MCP server that will ultimately host the API layer
What I’ll look for before sign-off:
• Langchain + LangGraph graphs executing end-to-end with live Azure AI Search indexes
• Reusable Python modules for GET/POST/PUT calls, fully typed and tested
• RAG responses that hit ≥90 % precision on our validation set
• A clear README explaining setup, env vars, and deployment via the MCP server
If you have shipped similar AI integrations and are comfortable debugging latency, token limits and embedding drift, this should feel straightforward. Let’s get the bot talking intelligently.
Every call the bot makes—GET, POST and PUT—must pass through a clean Python requests wrapper that you’ll write, document and unit-test. All code will live in a GitHub repo; I expect you to create concise branches, open pull-requests and keep the CI pipeline green.
What I’ll hand you:
• Current codebase with a minimal working chatbot
• Azure resources and keys
• Access to the MCP server that will ultimately host the API layer
What I’ll look for before sign-off:
• Langchain + LangGraph graphs executing end-to-end with live Azure AI Search indexes
• Reusable Python modules for GET/POST/PUT calls, fully typed and tested
• RAG responses that hit ≥90 % precision on our validation set
• A clear README explaining setup, env vars, and deployment via the MCP server
If you have shipped similar AI integrations and are comfortable debugging latency, token limits and embedding drift, this should feel straightforward. Let’s get the bot talking intelligently.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.